An Architecture for Data Unification in E-commerce using Graph

Abstract

Graph model has emerged as a contemporary technique for relationship-centric applications because of its various features like index-free adjacency, schema-less data, faster traversal of relationships, etc. In large-scale e-commerce applications, heterogeneous models are utilized for storing different types of data, e.g., relational model for transactional processing, ontology for product information, graph model for user preferences, etc. However, it creates overhead of accessing multiple data models for query processing. In this paper, we present a data modeling architecture for e-commerce which can be utilized for data unification from different data models to graph model. To verify the applicability of our approach, we analyze and compare query performance of our approach with heterogeneous data models. In addition, we also discuss the issues and challenges for adopting graph model for e-commerce applications.